Magnesium Modulates Actin Binding and ADP Release in Myosin Motors*

Background: Magnesium may be an important physiological regulator of myosin motor activity. Results: Mg2+ inhibits the ADP release rate constant in the subset of myosins examined and reduces actin affinity in the post-hydrolysis state in myosin V. Conclusion: Mg2+ alters contractile velocity without altering overall tension-generating capacity. Significance: Mg2+-dependent regulation of motor activity is conserved in myosin motors. We examined the magnesium dependence of five class II myosins, including fast skeletal muscle myosin, smooth muscle myosin, β-cardiac myosin (CMIIB), Dictyostelium myosin II (DdMII), and nonmuscle myosin IIA, as well as myosin V. We found that the myosins examined are inhibited in a Mg2+-dependent manner (0.3–9.0 mm free Mg2+) in both ATPase and motility assays, under conditions in which the ionic strength was held constant. We found that the ADP release rate constant is reduced by Mg2+ in myosin V, smooth muscle myosin, nonmuscle myosin IIA, CMIIB, and DdMII, although the ADP affinity is fairly insensitive to Mg2+ in fast skeletal muscle myosin, CMIIB, and DdMII. Single tryptophan probes in the switch I (Trp-239) and switch II (Trp-501) region of DdMII demonstrate these conserved regions of the active site are sensitive to Mg2+ coordination. Cardiac muscle fiber mechanic studies demonstrate cross-bridge attachment time is increased at higher Mg2+ concentrations, demonstrating that the ADP release rate constant is slowed by Mg2+ in the context of an activated muscle fiber. Direct measurements of phosphate release in myosin V demonstrate that Mg2+ reduces actin affinity in the M·ADP·Pi state, although it does not change the rate of phosphate release. Therefore, the Mg2+ inhibition of the actin-activated ATPase activity observed in class II myosins is likely the result of Mg2+-dependent alterations in actin binding. Overall, our results suggest that Mg2+ reduces the ADP release rate constant and rate of attachment to actin in both high and low duty ratio myosins.

We examined the magnesium dependence of five class II myosins, including fast skeletal muscle myosin, smooth muscle myosin, ␤-cardiac myosin (CMIIB), Dictyostelium myosin II (DdMII), and nonmuscle myosin IIA, as well as myosin V. We found that the myosins examined are inhibited in a Mg 2؉ -dependent manner (0.3-9.0 mM free Mg 2؉ ) in both ATPase and motility assays, under conditions in which the ionic strength was held constant. We found that the ADP release rate constant is reduced by Mg 2؉ in myosin V, smooth muscle myosin, nonmuscle myosin IIA, CMIIB, and DdMII, although the ADP affinity is fairly insensitive to Mg 2؉ in fast skeletal muscle myosin, CMIIB, and DdMII. Single tryptophan probes in the switch I (Trp-239) and switch II (Trp-501) region of DdMII demonstrate these conserved regions of the active site are sensitive to Mg 2؉ coordination. Cardiac muscle fiber mechanic studies demonstrate cross-bridge attachment time is increased at higher Mg 2؉ concentrations, demonstrating that the ADP release rate constant is slowed by Mg 2؉ in the context of an activated muscle fiber. Direct measurements of phosphate release in myosin V demonstrate that Mg 2؉ reduces actin affinity in the M⅐ADP⅐P i state, although it does not change the rate of phosphate release. Therefore, the Mg 2؉ inhibition of the actin-activated ATPase activity observed in class II myosins is likely the result of Mg 2؉dependent alterations in actin binding. Overall, our results suggest that Mg 2؉ reduces the ADP release rate constant and rate of attachment to actin in both high and low duty ratio myosins.
Magnesium is an abundant divalent cation known for its role as an essential metal cofactor for many enzymes that utilize nucleotide triphosphates (NTPs) as an intracellular energy source. Previous studies with motor proteins such as myosin traditionally focus on the role of Mg 2ϩ as a cofactor in ATP binding, hydrolysis, and product release in the force-generating mechanochemical cycle (1,2). Recent studies highlight the possibility that free Mg 2ϩ can modulate motor activity and thus physiological function in muscle myosin II and unconventional myosins I, V, and VII (3)(4)(5)(6)(7). These studies generally support a mechanism whereby Mg 2ϩ alters key steps in the mechanochemical cycle, including the ADP release rate constant, which is an important determinant of duty ratio (fraction of the ATPase cycle myosin remains attached to actin). However, it remains unclear to what degree other steps of the actomyosin ATPase cycle are impacted by Mg 2ϩ . In addition, it is unclear if Mg 2ϩ -dependent modulation of myosin motor activity is conserved throughout the myosin superfamily.
Several NTPase superfamilies, including myosins, kinesins, and G proteins, share highly conserved elements in their nucleotide binding regions (8). The P-loop, switch I, and switch II are involved in NTP binding, hydrolysis, and Mg 2ϩ coordination. In myosins, these conserved elements have the ability to amplify small changes in the nucleotide binding region to large scale force-generating movements of the light chain binding region or lever arm. Our previous work has focused on the impact of Mg 2ϩ on myosin V (MV), 2 a ubiquitous intracellular transporter that has been well characterized both structurally and biochemically (9). Studies in myosin V have supported a strain-sensitive two-state model for ADP release (10 -12). Myosin binding to actin in the post-hydrolysis ADP⅐P i state (K 9 ) results in acceleration of phosphate release (kЈ ϩ4 ) and forma-tion of the actomyosin⅐ADP state with a high affinity for ADP (Scheme 1). An active site isomerization results in a weakening of the affinity for ADP (kЈ ϩ5A ) and promotes the release of nucleotide (kЈ ϩ5B ). In the presence of actin, Mg 2ϩ is thought to exchange readily from both actomyosin⅐ADP states and can alter both the transition between actomyosin⅐ADP states and the release of ADP (5,6,9). Because the ADP release step is rate-limiting in myosin V, increasing free Mg 2ϩ concentration slows the myosin V ATPase activity as well as sliding velocity (9,13) based on a detachment-limited model for in vitro motility. The detachment-limited model (14) suggests that sliding velocity (V) is inversely proportional to the attached time (t on ), the period of time myosin is strongly attached to actin, and is proportional to the size of the unitary displacement (d uni ), the displacement generated by a single myosin motor (Equation 1).
Therefore, the Mg 2ϩ dependence of in vitro sliding velocity should correlate with the Mg 2ϩ dependence of the ADP release rate constant in myosins that follow a detachment-limited model.
Recent studies in other unconventional myosins support a conserved mechanism for Mg 2ϩ inhibition of the ADP release steps. In addition, these studies suggest that Mg 2ϩ can facilitate the transition between a motor that is a transporter or a tension generator (7). For example, regulation of myosin I occurs at Mg 2ϩ concentrations between 0.3 and 1 mM free Mg 2ϩ , which is within the concentration range measured intracellularly (4). A study on myosin VII, a mechanoenzyme involved in auditory and visual processes, reveals a similar sub-millimolar Mg 2ϩ sensitivity with total inhibition of sliding velocity above 1 mM free Mg 2ϩ (7). Although intracellular Mg 2ϩ concentrations are traditionally thought to be maintained within a narrow range, recent studies are challenging this view with the report of large Mg 2ϩ gradients in the cell (15). In muscle, the free Mg 2ϩ concentration ranges from 0.7 mM in cardiac muscle (16) to 0.9 -1.3 mM in skeletal muscle (17,18). In addition, alterations in free Mg 2ϩ concentrations have been associated with cardiovascular disease (19) and contractile performance (20). Taken together, these findings highlight the importance of investigating the conserved mechanisms of Mg 2ϩ -dependent modulation of myosin motor activity.
Structural and biochemical studies of Dictyostelium myosin II have contributed significantly to an enhanced understanding of the mechanochemical cycle of class II myosins (21). A study on Dictyostelium myosin II is in general agreement with myosin V studies regarding the presence of two actomyosin ADP states (3). However, this work proposed that Mg 2ϩ exchanges from the weak actomyosin⅐ADP state exclusively, which could potentially be the result of class-specific variations in Mg 2ϩ coordination. One residue in the switch II region of myosin V (Tyr-439 in myosin V, which corresponds to Ser-456 in Dictyo-stelium) was studied by Nagy et al. (22) and proposed to be an important mediator of Mg 2ϩ binding affinity. Variability at this conserved residue may be one contributing factor for the increased sensitivity to Mg 2ϩ observed in some unconventional myosins.
In this study, we directly compare the impact of Mg 2ϩ on a subset of muscle and nonmuscle myosins. We find the major impact of Mg 2ϩ manifests itself differently based on characteristic rate-limiting step differences in myosins. Our results support a model in which increasing Mg 2ϩ concentration universally slows actomyosin ATPase activity, in vitro actin sliding, and contractile velocity in muscle in a manner independent of ionic strength. We demonstrate that multiple steps in the actomyosin ATPase cycle are sensitive to changes in physiological free Mg 2ϩ concentrations.

EXPERIMENTAL PROCEDURES
Reagents and Actin Preparation-All reagents were of the highest quality and purity commercially obtainable. ATP and ADP were prepared fresh from powder, and the concentrations were measured by absorbance at 259 nm with ⑀ 259 ϭ 15,400 M Ϫ1 Ϫcm Ϫ1 (23). N-Methylanthraniloyl (mant)-labeled 2Ј-deoxy-ADP (mant-dADP) was purchased from Jena Scientific. The mant-dADP concentration was determined by absorbance at 255 nm and ⑀ 255 ϭ 23,300 M Ϫ1 ⅐cm Ϫ1 .
Protein Expression and Purification-The baculovirus/SF9 cell system was used to co-express a heavy meromyosin (HMM) version of Gallus gallus myosin VA containing an N-terminal FLAG tag and C-terminal YFP (MV) with calmodulin (24). A version of myosin VA containing a single IQ motif and C-terminal Myc and FLAG tags was co-expressed with calmodulin as described (25). The G. gallus smooth muscle myosin HMM with a C-terminal FLAG and Myc tag (SMII) (26) and Mus musculus nonmuscle myosin IIA HMM with a C-terminal FLAG tag (NMIIA) (27) were co-expressed with their associated light chains. All baculovirus/SF9 cell expressed constructs were purified by anti-FLAG affinity chromatography (24 -27). The single tryptophan containing (Trp-239 and Trp-501) Dictyostelium myosin II motor domain (DdMII) constructs was prepared as described (3,28). Rabbit psoas skeletal muscle was used to prepare skeletal muscle myosin HMM (SKII) (29). ␤-Cardiac HMM (CMIIB) was purified from pig hearts as described (29) with modification (30). Prior to characterization, SMII and NMIIA were phosphorylated with myosin light chain kinase (27). Rabbit smooth muscle light chain kinase was expressed in Sf9 cells, purified by anti-FLAG affinity chromatography, and stored at Ϫ80°C (24,26). An actin spin-down assay was performed on tissue-purified SKII and CMIIB prior to analysis to maximize the ratio of active to nonactive myosin heads. G. gallus calmodulin was expressed and purified as described (31). Myosin concentrations were determined by absorbance at 280 nm (MV 1IQ, ⑀ 280 ϭ 103,600 M Ϫ1 ⅐cm Ϫ1 , SKII, ⑀ 280 ϭ 210,000 M Ϫ1 ⅐cm Ϫ1 ) or by Bio-Rad microplate assay using bovine serum albumin (BSA) as a standard. Actin was prepared from rabbit skeletal muscle using the acetone powder method (32). Pyrene-labeled actin was prepared with N-(1-pyrene)iodoacetamide (Invitrogen) as described (33).   Fig. 1, in which 0.04 M was used) and 0.08 M for myosin V was maintained throughout all assays by varying KCl concentrations to balance changes in MgCl 2 concentration. The buffer used for DdMII experiments was 20 mM HEPES, pH 7.3, whereas MgCl 2 concentrations were varied, and ionic strength was compensated by changing NaCl concentrations (i.e. 120 mM NaCl, 0 mM MgCl 2 or 60 mM NaCl 2 , 20 mM MgCl 2 ). Free Mg 2ϩ concentrations were calculated using MaxC 2.5.

Magnesium-dependent Regulation of Myosin Motors
Stopped-flow Kinetics and Modeling-Most transient kinetic experiments were performed at 25°C in an Applied Photophysics stopped-flow spectrophotometer with a dead time of 1.2 ms. A monochromator with a 2-nm bandpass was used for fluorescence excitation, and a 395-nm cutoff filter was used to measure fluorescence emission. The mant-labeled nucleotides were excited by FRET from endogenous tryptophan residues (290 nm); pyrene actin was excited at 365 nm, and phosphate-binding protein was excited at 380 nm. Nonlinear least squares fitting of the data were conducted on software provided with the instrument or KaleidaGraph. The DdMII studies were performed in a Biokine (Biologic) stopped-flow apparatus, and tryptophan fluorescence was observed with an excitation wavelength of 297 nm, and emission was measured with a 340-nm interference filter. The DdMII stopped-flow experiment was performed at 13°C to maximize the tryptophan fluorescence signal. Reported uncertainties correspond to standard errors of the fits unless noted otherwise. Kinetic results were interpreted in terms of Schemes 1 and 2 with a description of the methods used to fit each specific experiment given under "Results" and figure legends.
ATPase Assays-Steady state ATPase activities were assessed using an NADH enzyme-linked assay at 25°C performed on the stopped flow (9,34,35). Absorbance changes were detected at 340 nm, and the extinction coefficient for NADH (⑀ 340 ϭ 6,220 M Ϫ1 ⅐cm Ϫ1 ) was used to determine the oxidation rate of NADH. Steady state ATPase rates were obtained at a set of MgCl 2 concentrations ranging from 0.5 to 10 mM (0.3 to 9 mM free Mg 2ϩ , respectively) in the presence of 20 M (MV) or 60 M (SKII, SMII, NMIIA, and CMIIB) actin and saturating 1 mM ATP. Assays were also performed with a range of actin concentrations to determine maximum ATPase rate (k cat ) and the actin concentration at which ATPase is half-maximal (K ATPase ).
In Vitro Motility and Filament Breaking Assays-The actin filament sliding assay was performed as described previously (36) at a range of Mg 2ϩ concentrations in conditions (controlled ionic strength) similar to the corresponding ATPase assay for each myosin. Myosin was adhered to a 1% nitrocellulose in an amyl acetate (Ladd Research)-coated coverslip directly or with a c-Myc antibody (Sigma) in the case of SMII and NMIIA. The surface was blocked with BSA at 1 mg⅐ml Ϫ1 before the addition of actin labeled with either rhodamine phalloidin (TRITC filter cube; excitation/emission, 545/620 nm) or Alexa (FITC filter cube; excitation/emission, 500/535 nm). An activation buffer with varying MgCl 2 concentrations (0.5-10 mM) was added containing MOPS or KMg50 with 5 M calmodulin (MV only) and supplemented with the corresponding amount of KCl to maintain ionic strength and the following: 0.35% methylcellulose, 2.5 mM phosphoenolpyruvate, 20 units⅐ml Ϫ1 pyruvate kinase, 0.1 mg⅐ml Ϫ1 glucose oxidase, 5 mg⅐ml Ϫ1 glucose, 0.018 mg⅐ml Ϫ1 catalase, and 1 mM ATP. The slide was promptly viewed using a NIKON TE2000 microscope equipped with a 60ϫ/1.4 NA phase objective and a Perfect Focus System. Images were acquired at intervals (0.2-5 s) and for periods of time (1-5 min) appropriate for each myosin using a shutter-controlled Coolsnap HQ2-cooled CCD digital camera (Photometrics) binned 2 ϫ 2. Temperature was maintained at 26 Ϯ 1°C and monitored using a thermocouple meter (Stable Systems International). Image stacks were transferred to ImageJ for analysis via MTrackJ (37). Frequency of filament breakage was determined for MV, NMIIA, and SKII at 0.3 and 9.0 mM free Mg 2ϩ by recording elapsed time to initial filament breakage for Ն50 filaments per condition (38).
Pig Myocardial Strips-All procedures were reviewed and approved by the Institutional Animal Care and Use Committees of the University of Vermont College of Medicine. Cardiac papillary muscle was taken from one male conventional swine (E. M. Parsons) weighing 16 kg. Myocardial strips were dissected further, chemically skinned, and studied at 27°C as described previously (39). Free Mg 2ϩ concentrations of 1, 2, 4, and 8 mM were applied at pCa 4.8 by exchanging volumes of 1 or 8 mM free Mg 2ϩ solutions.
Sinusoidal length perturbations of amplitude 0.125% muscle strip length were applied over 0.125-250 Hz. The elastic and viscous moduli were measured from the tension response, and Equation 2 was fitted to the complex modulus, where A, B, and C are magnitudes expressed in kilonewtons⅐m Ϫ2 ; k is a unit-less exponent; b and c are characteristic frequencies expressed in Hz; is angular frequency in units of s Ϫ1 of the length perturbation equal to 2 ϫ frequency of perturbation, and i ϭ (Ϫ1) 1/2 . The rate 2c is the equivalent of the myosin cross-bridge detachment rate, and its reciprocal is equal to time on, and the average time myosin is attached to actin (39).

RESULTS
To investigate the impact of Mg 2ϩ on actomyosin mechanochemistry, we examined the Mg 2ϩ dependence of a subset of myosins in three assays as follows: actin-activated ATPase, in vitro motility, and transient kinetic determination of the ADP release rate constant (Figs. 1-4). We examined a processive high duty ratio motor (MV), two class II myosins with relatively slow ADP release rate constants (NMIIA and SMII), and two low duty ratio muscle myosins (CMIIB and SKII). All assays were performed in ionic strength-controlled conditions by varying the concentration of KCl to offset the difference in MgCl 2 concentration.
Actin-activated ATPase Activity-Ionic strength-controlled steady state actin-activated ATPase assays for MV (Fig. 1A), SKII (Fig. 1B), SMII (Fig. 1C), and CMIIB ( Fig. 1D) were performed at 1 and 10 mM MgCl 2 (0.3 and 9.0 mM free Mg 2ϩ , respectively) and a range of actin concentrations. The analysis allowed for determination of the maximum actin-activated ATPase activity (k cat ) and the actin concentration at which the ATPase activity is half-maximal (K ATPase ) at low (0.3 mM) and high (9.0 mM) free Mg 2ϩ concentrations. SKII and SMII exhibited an approximate 2-fold decrease in k cat accompanied by a 2-fold increase in K ATPase at 9 mM free Mg 2ϩ (Table 1). MV and CMIIB exhibited a 2-and 6-fold decrease in k cat , respectively, and a 2-fold decrease in K ATPase at 9 mM free Mg 2ϩ . All class II myosins demonstrated a 3-fold reduction in k cat /K ATPase at high Mg 2ϩ , although MV was unchanged.
ADP Release Rate Constant-The Mg 2ϩ dependence of the ADP release rate constant was examined directly by monitoring the dissociation of mant-dADP in MV ( Fig. 2A), NMIIA (Fig.   2B), and SMII (Fig. 2C). A complex of actomyosin⅐mant-dADP was mixed with saturating ATP (final conditions: 0.75-1 M actin, 0.5-0.75 M myosin, 10 -20 M mant-dADP, and 1 mM ATP). The mant-dADP fluorescence transients followed a single exponential in SMII and a double exponential in MV and NMIIA as reported previously (24,26,27). The fast phase of the ADP release rate constants in MV and NMIIA was more steeply dependent on Mg 2ϩ concentration. The ADP release rate constants (fast phase for MV and NMIIA) were reduced ϳ2-fold at  Table 1 for a summary of k cat and K ATPase values.

TABLE 1
Summary of k cat (maximum ATPase rate) and K ATPase (actin concentration at half-maximal activation) values for actin-activated ATPase at 0.3 and 9 mM) free Mg 2؉ , determined by fitting the data in Fig. 1 to Michaelis-Menten kinetics 9.0 mM as compared with 0.3 mM free Mg 2ϩ and are shown for MV ( Fig. 2A, open red circles), NMIIA (Fig. 2B, open red circles), and SMII (Fig. 2C). The relative amplitudes of the double exponential fits with MV (83% fast phase) and NMMIIA (74% fast phase) were similar at all free Mg 2ϩ concentrations examined. To measure the ADP release rate constant in CMIIB, we examined ATP-induced dissociation from pyrene actin in the presence of ADP (Fig. 2D). Pyrene actin is quenched when myosin is strongly bound, and ATP binding allows examination of the rate of transition into the weakly bound states. A complex of pyrene actin⅐CMIIB in the presence of ADP was mixed with high concentrations of ATP, and the free Mg 2ϩ concentrations were varied (final conditions: 0.8 M pyrene actin and CMIIB, 160 M ADP, and 16 mM ATP). The transients were best fit by a three exponential function, and only the fast phase was dependent on free Mg 2ϩ . The relative amplitudes of the three phases were not Mg 2ϩ -dependent. Therefore, the fast phase was modeled to be the ADP release rate constant; the intermediate phase was modeled to be the previously identified slow isomerization in the nucleotide binding pocket (41), and the slow phase was a nonspecific fluorescence change. The ADP release rate constants were about 2-fold faster than previous measurements performed at 20°C and slightly different buffer conditions (41,42).
Because the ADP release rate constant is extremely fast in SKII, we measured the ADP affinity in CMIIB and SKII (Fig. 3, A and B and C and D, respectively). The ADP affinity at four different free Mg 2ϩ concentrations was measured by competition with ATP-induced dissociation monitored by pyrene actin fluorescence (43). At each free Mg 2ϩ concentration, the rate of ATP-induced dissociation was measured in the absence and presence of different concentrations of ADP. The transients (k obs ) were fit to a single exponential function, and the rate of ATP-induced dissociation decreased as a function of ADP concentration present. The rate constants were normalized to the rate in the absence of ADP (k rel ) (where k rel ϭ k obs /k 0 and k 0 ϭ k obs in the absence of ADP) and plotted as a function of ADP concentration. The k rel was plotted as a function of ADP concentration for CMIIB (Fig. 3, A and B) and SKII (Fig. 3, C and D). The data were fit to a previously established equation (k rel ϭ 1/(1 ϩ [ADP]/KЈ ADP )), where KЈ ADP is the apparent affinity for ADP (42), which was slightly enhanced at higher free Mg  (Fig. 4). The ADP release rate constant is the rate-limiting step in the MV solution ATPase assay, whereas attachment to actin or phosphate release are rate-limiting in the ATPase assay for class II myosins (23). We can therefore begin to elucidate which step(s) in the myosin ATPase cycle are impacted by Mg 2ϩ based on trends in the ATPase and in vitro motility assays, as well as the relative changes in the ADP release rate constant. The ATPase assays for myosin V and all four class II myosins were dependent on Mg 2ϩ ; however, MV was more steeply dependent on Mg 2ϩ . The in vitro motility sliding velocities exhibited similar magnesium dependences to the corresponding ATPase assay for all five myosins examined. The Mg 2ϩ dependence of the ADP release rate constant for MV, NMIIA, SMII, and CMIIB followed a similar trend as the in vitro motility assays. The ADP affinity of SKII (Fig. 4B), which was fairly insensitive to free Mg 2ϩ , was also plotted for comparison.
Kinetic Mechanism of Mg 2ϩ -dependent ADP Release in DdMII Motor-To further examine how Mg 2ϩ alters the kinetics of the myosin⅐ADP states, we performed transient kinetic studies with single tryptophan containing DdMII motor constructs. We examined the fluorescence changes in Trp-239 located in the switch I region, which decreases in fluorescence (15%) upon ADP binding. We also examined Trp-501 in the relay helix, which also decreases (15%) in fluorescence upon ADP binding. The rates of binding and release were monitored with both constructs. Similar values for the second-order rate constant (k ϩADP ) and the dissociation rate constant (k ϪADP ) were obtained with both constructs. The second-order rate constant for ADP binding to myosin was increased 10 -20-fold in the absence of Mg 2ϩ , although the dissociation rate constant was also increased 10-fold in the presence of saturating Mg 2ϩ (20 mM MgCl 2 ) ( Table 2 and Figs. 5 and 6). Our previous results (3) and the results in Fig. 7C are consistent with 20 mM MgCl 2 being saturating in DdMII. The rate of Mg 2ϩ binding to myosin⅐ADP was monitored with Trp-239 fluorescence (Fig. 7). We observed a fluorescence increase followed by a decrease upon Mg 2ϩ binding, and the rates and amplitudes were dependent on Mg 2ϩ concentration. Analysis of the data allowed for the determination of the rates of association, dissociation, and overall affinity. A summary of DdMII ADP states in the presence and absence of Mg 2ϩ satisfies the thermodynamic law of detailed balance (Scheme 2). Therefore, the overall affinity for ADP is relatively insensitive to Mg 2ϩ in DdMII, similar to what was observed in SKII and CMBII. The 10-fold faster rate constant for ADP dissociation in the absence of Mg 2ϩ is similar to what was observed in myosin V, although the impact on the second-order rate constant for ADP binding was less pronounced in myosin V (3-fold) (9). The basal ATPase activity is not altered by Mg 2ϩ in DdMII (0.03 s Ϫ1 at both 0.17 mM and 20 mM MgCl 2 ), which is consistent with Mg 2ϩ only altering the ADP release step in the absence of actin, which is not rate-limiting (44). . The fluorescence transients were fit to a single exponential function. The plots of phosphate release rate constant as a function of actin concentration were fit to a hyperbolic function to determine the affinity for actin in the M⅐ADP⅐P i state (1/K 9 ) and maximum rate of phosphate release (kЈ ϩ4 ) (Fig. 8). The kinetics of actin-activated phosphate release at 0.3 mM free Mg 2ϩ are similar to previous reports (1/K 9 ϭ 3.2 M, kЈ ϩ4 ϭ 106 s Ϫ1 ) (10,45). In the presence of 9 mM free Mg 2ϩ , the actin affinity in the M⅐ADP⅐P i state was 5-fold weaker (1/K 9 ϭ 16.8 M), although the rate of phosphate release was unchanged (kЈ ϩ4 ϭ 125 s Ϫ1 ).  Table 4 for velocity values). The data are expressed as relative values for comparison (normalized to the value obtained at 0.3 mM free Mg). The ADP release rate constant was examined with mant-dADP or pyrene actin (see Fig. 2), and ADP affinity was measured by competition with ATP-induced dissociation (see Fig. 3). AUGUST 22, 2014 • VOLUME 289 • NUMBER 34

JOURNAL OF BIOLOGICAL CHEMISTRY 23983
Filament Breaking Assay-An in vitro motility filamentbreaking assay was performed by measuring the average time to initial filament breakage in MV, NMIIA, and SKII (Table 3). Filament breaking analysis showed a 2-fold reduction in time to initial breakage at 9 mM free Mg 2ϩ compared with 0.3 mM free Mg 2ϩ in all three myosins. Although we cannot rule out that the increased free Mg 2ϩ alters the stability of the actin filaments, we found the actin filament lengths were similar at low and high free Mg 2ϩ (8 Ϯ 1 m) examined in the absence of ATP. The filament breaking results indicate that the attached time is increased at high free Mg 2ϩ concentrations, and the results parallel the relative reduction in ADP release rate constant observed for MV and NMIIA.
Pig Myocardial Mechanics and Kinetics- Fig. 9, A and B, illustrates the influence of varying free Mg 2ϩ concentrations on the elastic and viscous moduli, respectively. As the free Mg 2ϩ concentration was elevated from 2 to 8 mM, the elastic and viscous moduli shifted to lower frequencies. When the Mg 2ϩ concentration was then lowered to repeat the 2 mM condition, the moduli shifted back to higher frequencies, thus verifying reversibility of the effects.
The myosin cross-bridge detachment rate 2c was significantly lowered with increasing Mg 2ϩ concentrations (Fig. 9C). These data reflect a longer myosin cross-bridge attached time as a function of increased free Mg 2ϩ concentration (Fig. 9D).
Because ATP concentrations are saturating, the detachment rate measured here reflects the ADP release rate constant of the myosin cross-bridge. We did not detect any significant effect of Mg 2ϩ on developed tension (Fig. 9E). Our results here are qualitatively similar to those for skinned rat cardiac muscle reported by Puchert et al. (46), who report slower characteristics of force redevelopment as free Mg 2ϩ is raised from 0.5 to 1 mM. Our observations of the effects of free Mg 2ϩ concentration on cross-bridge ADP release rate constant in the skinned myocardial strip demonstrate that the findings in isolated pig cardiac myosin are detectable within the context of an intact myofilament lattice.

DISCUSSION
By examining the Mg 2ϩ dependence of a number of muscle and nonmuscle myosins, we demonstrate that conserved mechanisms likely mediate the impact of free Mg 2ϩ on the myosin mechanochemical cycle. It is well established that the ADP release rate constant plays a critical role in influencing the period of time myosin remains attached to actin in the presence of physiological ATP concentrations and therefore correlates with unloaded shortening velocity. Our current results demonstrate that the slowing of the ADP release rate constant as a function of increasing free Mg 2ϩ concentration is a conserved feature of myosin motors. We also find that Mg 2ϩ can alter the attachment of myosin to actin in the post-hydrolysis M⅐ADP⅐P i state, measured directly in myosin V. Our overall observations and analysis at the molecular level are consistent with our findings in skinned muscle fibers that demonstrate free Mg 2ϩ reduces the ADP release rate constant within the context of an intact myofilament lattice and would be expected to reduce contractile velocity of activated muscle while not altering overall muscle tension.
Impact on Detachment-limited Actomyosin Motility-We found that in the myosins examined, the Mg 2ϩ dependence of the in vitro motility sliding velocity and the ADP release rate constant followed a similar trend (Fig. 4). The relationship between ADP release and actomyosin motility has been well established (41,43,47). Our previous results demonstrated that the temperature dependence of in vitro motility also followed a similar trend to the temperature dependence of the ADP release rate constant (24). Overall, Mg 2ϩ dependence of in vitro sliding velocity can mostly be explained by a detachment-limited mechanism of motility. The detachment rate (1/t on ) at low and high Mg 2ϩ can be calculated using the in vitro sliding velocities (V) and the literature values of the working stroke (d uni ) (MV (48); SKII (49), SMII (50), NMII (51), and CMIIB (52)). In MV, NMIIA, CMIIB, and SMII, the calculated detachment rates are within 2-3-fold of the measured ADP release rate constants (Table 4). Differences between the calculated and measured detachment rates have been noted previously (51) and could result from a number of factors, including strain dependence of detachment or uncertainty in the working stroke values.
The release of ADP from actomyosin has been shown to be a two-step process in many myosins (10,26). The first step is the isomerization of the nucleotide binding pocket from a closed to an open conformation followed by local structural changes in the Scheme 2.

TABLE 2
Summary of rate and equilibrium constants for ADP binding measured with the single tryptophan containing DdMII constructs in the absence of MgCl 2 or in the presence of 20 mM MgCl 2 * 0.5 mM EDTA was used to chelate any residual Mg. a Data were determined from hyperbolic fit to data in Fig. 5, C and D. b Data were determined from hyperbolic fit to data in Fig. 6, C and D. c Data determined from hyperbolic fit to data in Figs. 5, E and F or 6, E and F. active site required for the release of ADP. We and others have found that both steps are Mg 2ϩ -dependent in myosin V (5,6,9) and that the isomerization step is rate-limiting in single-headed myosin V (9,34). The second step is also quite steeply dependent on Mg 2ϩ and may be the rate-limiting step in the myosin V dimer (53,54). We observed a single exponential transient in the mant-dADP release studies with SMII, whereas the fluorescence transients with MV and NMIIA were best fit by a twoexponential function. We associated the fast phase of the ADP release transient with the nucleotide release step and found that this rate correlates well with the Mg 2ϩ dependence of sliding velocity.
The ADP release kinetics in myosin V were more steeply dependent on free Mg 2ϩ than the class II myosins examined (NMIIA, SMII, and CMIIB). Structural differences in the coordination of Mg 2ϩ in the nucleotide binding pocket of class II and class V myosins may account for the differences observed in the Mg 2ϩ concentration-dependent inhibition of ADP release. The class II myosins examined all have a serine instead of a tyrosine in the switch II region that was demonstrated to be important 3, buffer at 13°C. Single exponential decays were fitted to the averages of 3-5 measurements. ADP concentration dependence of the fitted observed rates is shown for Trp-239 DdMII (C) and Trp-501 DdMII (D). The ADP binding to DdMII motor domain was treated kinetically as a second-order reaction. The observed rate constants fit a hyperbolic function of ADP concentration. The hyperbolic fit is indicated on each graph; C is the negative vertical asymptote; k ϩADP and k -ADP are the binding and unbinding rate constants, respectively. ADP concentration dependence of the amplitudes of the fluorescent transients is also shown for Trp-239 DdMII (E) and Trp-501 DdMII (F). Amplitudes were estimated from the offsets of the transients, because the dead time of the stopped-flow apparatus (ϳ0.3 ms) results in a considerable amplitude loss in the fits. K ADP , the ADP affinity of DdMII, is calculated from the hyperbolic ADP concentration dependence of the amplitudes. The calculated rate constants and ADP affinities from C, D, E, and F are summarized in Table 2. for Mg 2ϩ sensitivity (22). Our previous work demonstrated that the switch II region can play a role in mediating the transition between actomyosin⅐ADP states and the nucleotide release steps (55). The single tryptophan fluorescence experiments with DdMII demonstrate both switch I (Trp-239) and switch II (Trp-501) are sensitive to ADP binding (Figs. 5 and 6). The fluorescence of Trp-239 was sensitive to Mg 2ϩ binding in the myosin⅐ADP state (Fig.  7). Also, the relative amplitudes of the fluorescence changes observed in both Trp-239 and Trp-501 upon ADP binding were greater in the absence of Mg 2ϩ . Thus, our work highlights that both the switch I and switch II regions are critical for Mg 2ϩ coordination in the myosin⅐ADP states, and further mutational analysis may determine other key residues responsible for mediating differences in Mg 2ϩ affinity.
The ADP affinity measured in the striated muscle myosins (SKII and CMIIB) was relatively Mg 2ϩ -insensitive. One possible explanation for these results is that both the ADP binding and ADP dissociation rate constants are Mg 2ϩ -dependent, and thus the overall affinity is unchanged. The transient kinetic results with the single tryptophan containing DdMII constructs 3 buffer, at 13°C. Single exponential decays were fitted to the averages of 3-5 measurements. ADP concentration dependence of the fitted observed rates is shown for Trp-239 DdMII (C) and Trp-501 DdMII (D). The ADP binding to the DdMII motor domain was treated kinetically as a second-order reaction. The observed rate constants fit a hyperbolic function of ADP concentration. The hyperbolic fit is indicated on each graph; C is the negative vertical asymptote, and k ϩADP and k -ADP are the binding and unbinding rate constants; respectively. ADP concentration dependence of the fitted amplitudes of the fluorescent transients is also shown for Trp-239 DdMII (E) and Trp-501 DdMII (F). K ADP , the ADP affinity of DdMII is calculated from the hyperbolic ADP concentration dependence of the amplitudes. The calculated rate constants and ADP affinities from C, D, E, and F are summarized in Table 2. demonstrate that this is indeed the case for DdMII. The secondorder binding constant for ADP binding and the ADP release rate constant increase about 10 -20-fold in the absence of Mg 2ϩ , which results in relatively small changes in ADP affinity. However, the DdMII experiments were performed in the absence of actin, and the actin-bound conformations may alter Mg 2ϩ sensitivity. A similar trend was observed in myosin V in the presence of actin, although the second order binding constant was more modestly increased 2-3-fold in the absence of Mg 2ϩ (5,6,9).
Impact on Attachment to Actin-Our results demonstrate that the Mg 2ϩ dependence of in vitro sliding velocity follows a similar trend as relative actin-activated ATPase rates. These results were surprising because the two assays are thought to have different rate-limiting steps for all of the class II myosins examined. Our results with MV 1IQ demonstrate that the phosphate release rate constant is not altered by Mg 2ϩ , although the affinity for actin in the MV⅐ADP⅐P i state is 5-fold weaker. The results suggest a mechanism for why the class II myosins examined had a reduced k cat /K ATPase at higher Mg 2ϩ , because the rate-limiting step in these myosins is actin-activated phosphate release. We propose that Mg 2ϩ could alter attachment to actin by changing the electrostatic interactions known to be important for formation of the initial interaction between actin and myosin (56,57). Although the ionic strength   Our previous work demonstrated a rapid structural change in the actin binding region prior to phosphate release in myosin V (58), and this conformational change could be slowed by Mg 2ϩ binding to an allosteric site. The work of Baker and co-workers (38,59) suggests that myosins can follow attachment-limited motility under certain conditions (e.g. high ionic strength and low motor density). We considered the possibility that the impact of Mg 2ϩ on the attachment rate can play a role in altering sliding velocity in class II myosins. Baker and co-workers (38) have proposed a mechanism to explain how P i and blebbistatin inhibit both actin-activated ATPase and in vitro motility to a similar degree. Our results suggest the ADP release rate constant ( Fig. 4 and Table 4) plays a significant role in altering the Mg 2ϩ -induced changes in actin sliding velocity, whereas changes in actin attachment may also contribute.
Relationship between Strain Sensitivity and Mg 2ϩ -The impact of external load or strain on the ADP release step in myosins is well established (60 -63). Therefore, it is interesting to propose that that Mg 2ϩ sensitivity may have an impact on the load sensing properties of myosins. The structural mechanism of strain sensitivity is unknown, although the conformation of the force-sensing lever arm may be allosterically coupled to structural changes in the nucleotide binding pocket that are associated with ADP binding. The coordination of Mg 2ϩ in the FIGURE 9. Elastic and viscous moduli measured at maximum calcium activation and varying free Mg 2؉ concentrations. A and B, characteristic dips and peaks of the elastic and viscous moduli were shifted to lower frequencies as free Mg 2ϩ concentration was elevated. The repeat measurement at 2 mM free Mg 2ϩ indicates that these effects were reversible. C, myosin cross-bridge detachment rate was detected by analysis of the frequency characteristics of the elastic and viscous moduli and was found to be reduced with increasing free Mg 2ϩ concentrations. D, corresponding cross-bridge time-on was prolonged with increasing free Mg 2ϩ concentrations. E, developed isometric tension was not significantly affected by Mg 2ϩ . One-way analysis of variance demonstrated significance at p Ͻ 0.01 level for detachment rate and time on. rep ϭ repeat. *, p Ͻ 0.05; **, p Ͻ 0.01 compared with 1 mM condition, and †,p Ͻ 0.01 compared with 8 mM condition by paired t test. n ϭ 4 strips. a Sliding velocities were determined in the in vitro motility assay in the presence of low free Mg (0.3 mM for all except CMIIB, which is shown at 1.1 mM free Mg for comparison with ADP release) and high free Mg (9 mM). b Data were determined by (1/t on ) ϭ V/d uni , where d uni was taken from the literature values referenced in the text. c Data were determined from the mant-dADP release studies in MV (fast phase), NMIIA (fast phase), SMII (free Mg is the same as velocity and 1/t on ), and pyrene-actin experiments in CMIIB (low free Mg ϭ 1.5 mM and high free Mg ϭ 9.5 mM).
active site may be part of this allosteric coupling pathway. The switch II region has been proposed to play a role in communications between the active site and the lever arm (64). Our previous results that examined mutations in the switch II region of myosin V (55) and the current results demonstrating the Mg 2ϩinduced fluorescence changes in DdMII Trp-501 suggest Mg 2ϩ coordination by switch II may be an important part of this mechanism. Further investigation is necessary to elucidate the structural details of the strain-sensitive ADP release mechanism and the role of Mg 2ϩ in this process. Impact on Muscle Function-The cardiac skinned muscle fiber studies clearly demonstrate that attachment time is increased at higher Mg 2ϩ concentrations. Thus, in the context of an actively contracting muscle fiber, it is likely that higher Mg 2ϩ concentrations slow the ADP release rate constant and reduce contractile velocity. Our results exhibiting unchanged muscle tension as a function of Mg 2ϩ support a mechanism by which the rate of attachment is also altered by Mg 2ϩ . We propose the rate of attachment and detachment are both altered in the muscle fiber studies, which leads to little change in muscle tension. We would speculate that the overall effect of a reduced detachment rate and preserved tension with increasing Mg 2ϩ on muscle performance would include a reduction in maximum contractile power (tension ϫ velocity) in proportion to the reduction in maximum velocity. The optimum velocity, i.e. the velocity at which maximum power is produced, would be also significantly reduced, but the optimum tension would not change. Because the lifetime of the actomyosin⅐ADP state is load-dependent in both muscle (60) and nonmuscle myosins (61,63), Mg 2ϩ inhibition of contractile velocity may be more pronounced at high loads. The sensitivity of these and other attributes of muscle performance to free Mg 2ϩ warrant further investigation.
Our previous work demonstrated a correlation in the actinactivated ATPase activity and in vitro motility in four different myosins (MV, SKII, SMII, and NMIIA) (24). A similar relationship was revealed in a landmark study by Barany (65) in that the contractile velocity correlated well with maximal ATPase rate in myosins isolated from a variety of muscles. However, as pointed out by White and co-workers (47), a correlation between actin-activated ATPase rate and sliding velocity does not confirm that the two parameters have the same rate-limiting step. Indeed, experimental results suggest that contractile velocity is limited by detachment from actin, which agrees well with the ADP release rate constant (24,47), although the ATPase rate is limited by binding to actin and/or phosphate release and is too slow to limit sliding velocity. Our results highlight that Mg 2ϩ and temperature can both alter multiple steps in the ATPase cycle, which results in specific changes in contractile velocity and force generation.
We cannot rule out the possibility that Mg 2ϩ alters the function of regulatory proteins in muscle, such as troponin C, and the regulatory light chain by competing with calcium-binding sites. One study has addressed this question by examining the Mg 2ϩ dependence of rat cardiac muscle with and without hyperthyroid treatment, which shifts the myosin isoform expression in the ventricles from ␣to ␤-cardiac myosin (46). Because the Mg 2ϩ dependence of cardiac muscle was more dra-matic when ␣-cardiac myosin was expressed, without a change in myosin light chain expression, they concluded that the impact of Mg 2ϩ was associated with its effects on the myosin motor.
Changes in free Mg 2ϩ or ADP concentrations will shift the MgADP/ADP ratio causing an impact on myosin motor activity that is limited by the ADP release rate constant (e.g. contractile speed and/or intracellular transport). There have been reports that the free Mg 2ϩ concentrations in cardiac muscle increase in ischemic heart muscle and thus could play a role in the pathophysiology of heart function (66,67). Because ischemic episodes are often followed by prolonged hypercontraction without effective relaxation, the Mg 2ϩ -induced increase in the myosin cross-bridge lifetime may play a role in this and other pathological conditions. The free Mg 2ϩ concentrations increase in skeletal muscle during exercise and initial recovery (68), which suggests that Mg 2ϩ could play a role in muscle fatigue. Mg 2ϩ concentrations in the brain were found to vary in different neurological disorders (69) and thus could alter the function of myosin motors, such as myosin V, in neurons. A transient influx of Mg 2ϩ was found to play a critical role in T-cell activation, and a mutation in the Mg 2ϩ transporter associated with this process results in immunodeficiency disease (70). Further studies are necessary to establish the impact of Mg 2ϩ in mediating in vivo myosin motor function in different pathological and physiological conditions.
Conclusions and Future Directions-We have found that Mg 2ϩ alters myosin motor activity by altering steps associated with actin attachment and detachment. The mechanism of altering detachment is likely due to Mg 2ϩ exchange in the active site, although it is unclear how Mg 2ϩ impacts attachment. Future studies will examine how Mg 2ϩ can alter the physiological function of muscle and nonmuscle myosins.